Optimization algorithms and
software. Faster and more effective algorithms and software for
nonlinear, mixed-integer, and linear programming.
Feasibility and infeasibility in
optimization. Ways of reaching a feasible solution more quickly
for nonlinear and mixed-integer programs, and of analyzing infeasible
optimization models. Spin-off applications from algorithms for
assistants. Automated tools for analyzing and debugging
optimization models. For example, one tool analyzes the shape of
nonlinear functions and regions to help select the correct solver.
Applied optimization. Examples
include transistor sizing, DSP task-to-processor assignment, flexible
manufacturing systems, forestry, scheduling, task assignment in cloud
computing, channel assignment in wireless networks, 3G communications
Data classifiers. A new
approach for finding good data classifiers arises from an infeasibility
analysis algorithm. What is the best way to use this to develop
better data classifiers?